The goal of this project is to elucidate structure-function relationships in macromolecular machines. During FY15, our studies focussed on a junctional protein in the human retina; a murine transposase; further characterization of an iron-sequestering bacterial nano-compartment; and a minor contribution to an investigation of the STING protein which is implicated in early onset systemic inflammation, vasculitis, and pulmonary inflammation. (i) Back-to-back octameric rings of the retiinoschisin protein (RS1) suggest a junctional model for cell adhesion in the retina. Retinoschisin (RS1) is a cell adhesion protein required to maintain the structural and functional integrity of the retina. Mutations in RS1 leads to early vision impairment in young males, X-linked retinoschisis (XLRS), characterized by separation of inner retinal layers and disrupting synaptic signaling. From prior work, RS1 has been thought to form an octamer, with each subunit comprising a discoidin domain and a small N-terminal (RS1) domain. We have used cryo-electron microscopy to determine the structure of RS1 at 0.4 nm which is in fact a complex of two opposing octameric rings. In a ring, each subunit has the canonical discoidin fold. The RS1 domains occupy the centers of the rings, but are less clearly defined, suggesting mobility. We modeled the DS domains, consistent with intramolecular and intermolecular disulfides previously reported. The interfaces internal to and between rings feature residues implicated in XLRS, indicating the importance of correct assembly. Loops at the periphery of the rings may bind sugars and lipids on the membrane surface. As RS1 is entirely extracellular and without membrane-embedded loops, it apparently couples neighboring retinal membranes together through octamer-octamer contacts, perhaps modulated by interactions with other membrane components. These observations have been submitted for publication. (ii) The Hermes protein is a member of the hAT transposon superfamily which has active representatives, including McClintock's archetypal Ac mobile genetic element, in many eukaryotic species. Our colleagues determined the crystal structure of the Hermes transposase-DNA complex which revealed that Hermes forms an octamer, and that each monomer has bound a cleaved transposon end. We contributed EM analyses that localized the BED domains that were invisible, i.e. poorly ordered, in the crystal (3). The overall picture is that the catalytic unit is a dimer: however, only octamers are active in vivo. This suggests that they provide crucial multiple specific DNA binding domains that recognize repeated subterminal sequences and non-specific DNA binding surfaces for target capture. The unusual structure explains the basis of bipartite DNA recognition at hAT transposon ends, provides a rationale for transposon end asymmetry, and demonstrates how an octamer could provide multiple sites of interaction to allow the transposase to locate its transposon ends amidst a sea of chromosomal DNA. Our negative staining analysis was included in a paper published in the past year (1) that presents the results outlined above and is being followed up with exploratory cry-EM experiments on Hermes/DNA complexes. (iii) A virus capsid-like nanocompartment that stores iron and protects bacteria from oxidative stress. Living cells compartmentalize materials and enzymatic reactions to increase their metabolic efficiency. While eukaryotes use membrane-bound organelles, bacteria and archaea rely primarily on protein-bound nanocompartments. Encapsulins are a recently discovered class of nanocompartments. Hitherto, their functions have been unclear. We have characterized the structure of the encapsulin nanocompartment from Myxococcus xanthus and shown that its role is to sequester cytosolic iron, thereby to protect the cells from oxidative stress. This nanocompartment consists of a protein shell with internal contents. It has a shell protein (EncA, 32.5 kDa) and three internal proteins (EncB, 17 kDa; EncC, 13 kDa; EncD, 11 kDa). Using cryo-EM, we determined that EncA expressed in E. coli self-assembles into an icosahedral shell 32 nm in outer diameter built from 180 subunits with the fold first observed in bacteriophage HK97 capsid. Native nanocompartments have dense iron-rich cores. Functionally, they resemble ferritins, cage-like iron storage proteins, but with a massively greater capacity (30,000 Fe atoms vs. 3,000 in ferritin). Physiological data reveal that few nanocompartments are assembled during vegetative growth, but they increase five-fold upon starvation, protecting cells from oxidative stress through iron sequestration. Since publishing the results outlined above in FY14 (3), we are extending the investigation in three directions: seeking to determine the crystal lattice spacings in the iron-phosphate granules by electron diffraction, to identify the crystal form of mineralization; expressing the internal proteins with an aim to crystallographic sides; and seeking to extend the resolution of cryo-EM analyses. . (iv) Molecular modeling of mutated STING residues in clinically affected children. In a multi-participant multi-faceted study by a consortium led by Dr Raphaela Goldbach-Mansky (NIAMS), a cohort of 6 patients was identified with early onset systemic inflammation, vasculitis, and pulmonary inflammation. These patients were found to have de novo gain-of-function mutations in TMEM173, which encodes STimulator of Interferon Genes (STING). We participated in this effort by using molecular graphics and modeling to map the mutated residues on a homology model of a STING dimer. Intriguingly, the mutations tend to map at or near the dimer interface (2). (v) All cells must be capable of degrading aberrant and foreign proteins that would otherwise pollute them. These activities are carried out by energy-dependent proteolytic machines, which consist of two subcomplexes - a protease and an ATPase/unfoldase. Since 1995, we have studied the Clp complexes of E. coli, considered as a model system. We described the structures of the two sub complexes and characterized the interactions between them and with bound substrate proteins. In FY15, this project was largely dormant but it is still in our portfolio.